MRF B1 Correction Using Linear Dictionary Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic Resonance Fingerprinting (MRF) techniques face computational complexity and precision issues due to B1 inhomogeneity, particularly at high flip angles and ultrahigh fields, leading to biased quantification of tissue properties.
Innovation Solution
A linear mapping approach is used to correct B1 inhomogeneity by generating a nominal MRF dictionary and subsampling it to create corrected dictionaries, with computed transformation matrices for efficient B1 correction, reducing computational complexity and memory requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If B1 inhomogeneity is accounted for by building varied B1 values into the MRF dictionary, then measurement precision of tissue properties is improved, but device complexity and memory requirements increase exponentially
Solution Approach 1:
The patent segments the B1 correction problem by separating the nominal dictionary (B1=1) from B1 variation correction. Instead of including all possible B1 values in one large dictionary, the method divides the correction into: (1) a base nominal dictionary with B1=1, and (2) separate B1 variation correction terms that are applied multiplicatively. This segmentation reduces the base dictionary size while maintaining correction capability through separate correction factors.
Solution Approach 2:
The patent applies parameter changes by transforming the dictionary entries through B1 correction factors rather than storing all possible B1 values directly. The correction is implemented by multiplying nominal dictionary entries by B1 correction factors derived from measured B1 maps, effectively changing the parameter representation from absolute B1 values to relative correction factors, thereby reducing dictionary size while maintaining precision.
2Measurement precision
If a full MRF dictionary with multiple B1 values is constructed, then tissue property estimation accuracy is improved, but computational time and processing speed decrease
Solution Approach 1:
The patent applies preliminary action by pre-computing B1 correction factors from B1 maps before the main pattern matching process. These correction factors are stored and applied during reconstruction, avoiding the need to search through a large dictionary containing all possible B1 variations. This preliminary computation of correction terms speeds up the main processing while maintaining accuracy.
Solution Approach 2:
The patent extracts the B1 correction information from the main dictionary search process. Instead of searching through a large dictionary that includes varied B1 values, the method extracts B1 correction factors separately and applies them as multiplicative corrections to nominal dictionary entries. This extraction separates the correction function from the search function, improving processing speed while maintaining precision.
3Measurement precision
If B1 correction is performed pixel-by-pixel using a large dictionary with varied B1 values, then measurement precision is improved, but loss of time increases due to separate processing for each pixel
Solution Approach 1:
The patent merges the B1 correction operation with the nominal dictionary lookup operation. Instead of performing separate correction steps for each pixel using a large dictionary, the method combines the correction by multiplying the nominal dictionary entries by pre-computed B1 correction factors. This merging of operations allows simultaneous completion of lookup and correction, reducing processing time while maintaining pixel-by-pixel precision.
Data Source
AI summary
A magnetic resonance fingerprint (MRF) method includes obtaining an MRF undersampled image, and determining a corresponding radiofrequency transmit magnetic field strength for a pixel of the image. The method also includes obtaining a tissue property from an MRF dictionary based on the pixel, and correcting the tissue property based on the determined radiofrequency transmit magnetic field strength. According to another aspect, a MRF method includes generating a full MRF dictionary based on a nominal radiofrequency transmit magnetic field strength and, for each of a plurality of radiofrequency transmit magnetic field strengths different than the nominal field strength, subsampling the full MRF dictionary and generating a corrected MRF dictionary based on the subsampled dictionary and the different field strengths. The method also includes determining a transformation matrix between the full MRF dictionary and each of the corrected MRF dictionaries.